﻿Title		Replication Data, Key Analyses (3-26-20).
Get File = "C:\Users\ricklau\Box\Documents\Andersen\Multiday\Data\MultiDayReplication.sav".

Subtitle	Codebook --  Table A-1.
Descriptives	All.

*Comment	************************************************.

Subtitle	FIGURE 4, Discretionary Attention to Issues.
Frequencies	GHPolBrutTOD,GHPFLTOD,GHImmigTOD,GHFlagBurnTOD/
		Format = Notable/Histogram Normal.


*Comment	************************************************.
*Comment	************************************************.

Title		POLICE BRUTALITY ISSUE.
Subtitle	Police Brutality: Crucial Variables.
Descriptives	D1PolBrut,EDPolBrut,
		PolBrutCount,PolBrutDaysWt,PolBrutRecent,
		PolBrutDays,PolBrutRecentReceived,
		PolBrutOpens,PolBrutUniqueOpens, 
		GHPolBrutAtt,GHSPolBrutAtt,GHPolBrutTOD,
		D1PolBrutImp,EDPolBrutImp.

Subtitle	Agenda-Setting, Police Brutality: ITT (Table A-3, top).
Regression	Variables = EDPolBrutImp,PolBrutCount,
		Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,Dem/Depend = EDpolBrutImp/Enter.
*NS, neg.

Subtitle	Police Brutality: ITT, Most Recent (Table A-3, bottom).
Regression	Variables = EDPolBrutImp,PolBrutRecent,
		Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,Dem/Depend = EDpolBrutImp/Enter.
*NS, neg.

*Comment	*****************************************
		Assumes Agenda-Setting = Opportunity.

Subtitle        1st Stage Using Manip to Predict Good PolBrutDays-Table A-5.
Regression	Variables = PolBrutDays,PolBrutCount,
	 	Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Dem/
		Depend = PolBrutDays/Enter/Save = Pred(PredPolBrutCount).
Regression	Variables = PolBrutRecentReceived,PolBrutRecent,
	 	Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Dem/
		Depend = PolBrutRecentReceived/Enter/
		Save = Pred(PredPolBrutRecent).
*Note: Very good instruments!.
Variable Labels	PredPolBrutCount "Instrument for PolBrutDays"/
		PredPolBrutRecent "Instrument for PolBrutRecentReceived".
Descriptives	PredPolBrutCount,PredPolBrutRecent.

*Comment	*****************************************
		Assumes Agenda-Setting = Opportunity.

Subtitle	Now 2SLS, Estimating CACE (Table 1).
2SLS		EDPolBrutImp with PolBrutDays, Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon,Dem/Instruments = PolBrutCount,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Dem.
*NOTE: Treatment negative, ns.

Subtitle	2SLS, Estimating Interaction CACE (Table 2).
Compute		Int = Dem*PredPolBrutCount.
2SLS		EDPolBrutImp with PolBrutDays, Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon,Dem,Int/Instruments = PolBrutCount,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Dem,Int.
*NS int.

Subtitle	Separate Groups Procedure for Estimating Crucial Interaction.
temporary.
Select if	(Dem EQ 0).
2SLS		EDPolBrutImp with PolBrutDays, Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon/Instruments = PolBrutCount,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon.
temporary.
Select if	(Dem EQ 1).
2SLS		EDPolBrutImp with PolBrutDays, Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon/Instruments = PolBrutCount,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon.
*NEWNOTE:  Same basic story: in-party estimates increase from -.193 to -.054.


Subtitle	2SLS, Estimating CACE, Most Recent (Table A-7, top).
2SLS		EDPolBrutImp with PolBrutRecentReceived, Female,NonWhiteH,
		Age,EducYrs,Income,PolInterest,Libcon,Dem/
		Instruments = PolBrutRecent, Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon,Dem.
*ns., neg.

Subtitle	2SLS, Estimating Interaction CACE (Table A-7, bottom).
Compute		Int = Dem*PredPolBrutRecent.
2SLS		EDPolBrutImp with PolBrutRecentReceived, Female,NonWhiteH,
		Age,EducYrs,Income,PolInterest,Libcon,Dem,Int/
		Instruments = PolBrutRecent, Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon,Dem,Int.
*NS, Int, negative.

*Comment	*********************************************
		In the Broadcast News era, agenda setting involved 
		actually LOOKING AT the news stories; This is now 
		correlation world, but, Does actual (discretionary) 
		search behavior predict Change in the perceived 
		importance of the manipulated issue?.

*Comment        ****************************************.
 
Subtitle	Does Info Search Predict Perceived Importance? (Table A-2).
Regression	Variables = EDPolBrutImp,GHPolBrutTOD,
		Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,Dem/Depend = EDPolBrutImp/Enter.
*ns, and small negative.

*Comment	****************************************
		Assumes Agenda-Setting = Stories read.

Subtitle	First Stage, Using Manip to Predict Discretionary Attention.
Regression 	Variables = GHPolBrutTOD,PolBrutCount, D1PolBrut,
	 	Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Dem/
		Depend = GHPolBrutTOD/Enter/Save = Pred(PredPolBrutCountTOD).
Regression 	Variables = GHPolBrutTOD,PolBrutRecent, D1PolBrut,
	 	Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Dem/
		Depend = GHPolBrutTOD/Enter/Save = Pred(PredPolBrutRecentTOD).
*NOTE: Instruments only mediocre, with noticeably shrinking variance.
Descriptives	PredPolBrutCountTOD,PredPolBrutRecentTOD.

*Comment	****************************************
		Assumes Agenda-Setting = Stories read.

Subtitle	Now 2SLS Again, Re-Estimating CACE (TOD rather than count).
2SLS		EDPolBrutImp with GHPolBrutTOD,Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon,Dem/
		Instruments = PolBrutCount, D1PolBrut,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Dem.
*NS, negative.

Compute		Int = Dem*PredPolBrutCountTOD.
2SLS		EDPolBrutImp with GHPolBrutTOD,Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon,Dem,Int/
		Instruments = PolBrutCount, D1PolBrut,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Dem,Int.
*Int positive, but trivial.

2SLS		EDPolBrutImp with GHPolBrutTOD,Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon,Dem/
		Instruments = PolBrutRecent, D1PolBrut,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Dem.
*NS, negative.

Compute		Int = Dem*PredPolBrutRecentTOD.
2SLS		EDPolBrutImp with GHPolBrutTOD,Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon,Dem,Int/
		Instruments = PolBrutRecent, D1PolBrut,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Dem,Int.
*NS, int negative.


*Comment	************************************************.
*Comment	************************************************.

Title		PAID FAMILY LEAVE.
Subtitle	Paid Family Leave: Crucial Variables.
Descriptives	D1PFL,EDPFL,
		PFLCount,PFLDaysWt,PFLRecent, 
		PFLDays,PFLRecentReceived,
		PFLOpens,PFLUniqueOpens,
		GHPFLAtt,GHSPFLAtt,GHPFLTOD,
		D1PFLImp,EDPFLImp.

*Comment        **********************************************
                Does Assigned Treatment Affect Agenda?.

Subtitle	Agenda-Setting, Paid Family Leave: ITT (Table A-3, top).
Regression	Variables = EDPFLImp,PFLCount,
		Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,Dem/Depend = EDPFLImp/Enter.
*NS, at least positive.

Subtitle	Paid Family Leave: ITT, Most Recent (Table A-3, bottom).
Regression	Variables = EDPFLImp,PFLRecent,
		Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,Dem/Depend = EDPFLImp/Enter.
*ns, positive.

*Comment	*****************************************
		Assumes Agenda-Setting = Opportunity.

Subtitle        First Stage,Using Manip to Predict PFLDaysGood.  (Table A-5).
Regression	Variables = PFLDays,PFLCount,
	 	Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Dem/
		Depend = PFLDays/Enter/Save = Pred(PredPFLCount).
Regression	Variables = PFLRecentReceived,PFLRecent,
	 	Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Dem/
		Depend = PFLRecentReceived/Enter/
		Save = Pred(PredPFLRecent).
*Note: Excellent instruments!.
Variable Labels	PredPFLCount "Instrument for PFLDays"/
		PredPFLRecent "Instrument for PFLRecentReceived".
Descriptives	PredPFLCount,PredPFLRecent.

*Comment	*****************************************
		Assumes Agenda-Setting = Opportunity.

Subtitle	Now 2SLS, Estimating CACE (Table 1).
2SLS		EDPFLImp with PFLDays, Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,Dem/Instruments = PFLCount,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Dem.
*ns, but positive.

Subtitle	2SLS, Estimating Interaction CACE (Table 2).
Compute		Int = Dem*PredPFLCount.
2SLS		EDPFLImp with PFLDays, Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,Dem,Int/Instruments = PFLCount,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Dem,Int.
*NS int.

Subtitle	Separate Groups Procedure for Estimating Crucial Interaction.
temporary.
Select if	(Dem EQ 0).
2SLS		EDPFLImp with PFLDays, Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon/Instruments = PFLCount,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon.
temporary.
Select if	(Dem EQ 1).
2SLS		EDPFLImp with PFLDays, Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon/Instruments = PFLCount,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon.
*NEWNOTE: Same basic story: treatment effect increases from .111 to .264.


Subtitle	2SLS, Estimating CACE, Most Recent (Table A-7, top).
2SLS		EDPFLImp with PFLRecentReceived, Female,NonWhiteH,
		Age,EducYrs,Income,PolInterest,Libcon,Dem/
		Instruments = PFLRecent, Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon,Dem.
*Close! P < .14.

Subtitle	2SLS, Estimating Interaction CACE (Table A-7, bottom).
Compute		Int = Dem*PredPFLRecent.
2SLS		EDPFLImp with PFLRecentReceived, Female,NonWhiteH,
		Age,EducYrs,Income,PolInterest,Libcon,Dem,Int/
		Instruments = PFLRecent, Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon,Dem,Int.
*NS Int, negative.

*Comment        ****************************************.

Subtitle	Does Info Search Predict Perceived Importance? (Table A-2).
Regression	Variables = EDPFLImp,GHPFLTOD,
		Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,Dem/Depend = EDPFLImp/Enter.
*Close!: p < .06.

*Comment	****************************************.
*Comment	****************************************
		Assumes Agenda-Setting = Stories read.

Subtitle	First Stage, Using Manip to Predict Discretionary Attention.
Regression 	Variables = GHPFLTOD,PFLCount, D1PFL,
	 	Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Dem/
		Depend = GHPFLTOD/Enter/Save = Pred(PredPFLCountTOD).
Regression 	Variables = GHPFLTOD,PFLRecent, D1PFL,
	 	Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Dem/
		Depend = GHPFLTOD/Enter/Save = Pred(PredPFLRecentTOD).
*NOTE: Passable instruments, but restricted variance.
Descriptives	PredPFLCountTOD,PredPFLRecentTOD.

*Comment	****************************************.
*Comment	****************************************
		Assumes Agenda-Setting = Stories read.

Subtitle	Now 2SLS Again, Re-Estimating CACE.
2SLS		EDPFLImp with GHPFLTOD, Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon,Dem/
		Instruments = PFLCount, D1PFL,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Dem.
*Close! p < .06.

Compute		Int = Dem*PredPFLCountTOD.
2SLS		EDPFLImp with GHPFLTOD, Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon,Dem,Int/
		Instruments = PFLCount, D1PFL,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Dem,Int.
*No Int.

2SLS		EDPFLImp with GHPFLTOD, Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon,Dem/
		Instruments = PFLRecent, D1PFL,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Dem.
*Sig! p < .04.

Compute		Int = Dem*PredPFLRecentTOD.
2SLS		EDPFLImp with GHPFLTOD, Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon,Dem,Int/
		Instruments = PFLRecent, D1PFL,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Dem,Int.
*NS, int negative.


*Comment        **********************************************.
*Comment        **********************************************.

Title		IMMIGRATION REFORM/RESTRICTION.
Subtitle	Immigration Reform: Crucial Variables.
Descriptives	D1Immig,EDImmig,
		ImmigCount,ImmigDaysWt,ImmigRecent,
		ImmigDays,ImmigRecentReceived,
		ImmigOpens,ImmigUniqueOpens, 
		GHImmigAtt,GHSImmigAtt,GHImmigTOD,
		D1ImmigImp,EDImmigImp.

*Comment        **********************************************.
*Comment        **********************************************
		Does Assigned Treatment Affect Agenda?.

Subtitle	Agenda-Setting, Immigration Reform: ITT (Table A-3, top).
Regression	Variables = EDImmigImp,ImmigCount,
		Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,Rep/Depend = EDImmigImp/Enter.
*NS, negative.

Subtitle	Immigration Reform: ITT, Most Recent (Table A-3, bottom).
Regression	Variables = EDImmigImp,ImmigRecent,
		Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,Rep/Depend = EDImmigImp/Enter.
*NS.

*Comment	*****************************************
		Assumes Agenda-Setting = Opportunity.

Subtitle        1st Stage, Using Manip to Predict Good ImmigDays (Table A-5).
Regression	Variables = ImmigDays,ImmigCount,
	 	Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Rep/
		Depend = ImmigDays/Enter/Save = Pred(PredImmigCount).
Regression	Variables = ImmigRecentReceived,ImmigRecent,
	 	Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Rep/
		Depend = ImmigRecentReceived/Enter/
		Save = Pred(PredImmigRecent).
*Note: Very good instruments!.
Variable Labels	PredImmigCount "Instrument for ImmigDays"/
		PredImmigRecent "Instrument for ImmigRecentReceived".
Descriptives	PredImmigCount,PredImmigRecent.

*Comment	*****************************************
		Assumes Agenda-Setting = Opportunity.

Subtitle	Now 2SLS, Estimating CACE (Table 1).
2SLS		EDImmigImp with ImmigDays, Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon,Rep/Instruments = ImmigCount,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Rep.
*NS, negative.

Subtitle	2SLS, Estimating Interaction CACE (Table 2).
Compute		Int = Rep*PredImmigCount.
2SLS		EDImmigImp with ImmigDays, Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon,Rep,Int/Instruments = ImmigCount,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Rep,Int.
*NS, int negative.

Subtitle	Separate Groups Procedure for Estimating Crucial Interaction.
temporary.
Select if	(Rep EQ 0).
2SLS		EDImmigImp with ImmigDays, Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon/Instruments = ImmigCount,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon.
temporary.
Select if	(Rep EQ 1).
2SLS		EDImmigImp with ImmigDays, Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon/Instruments = ImmigCount,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon.
*NEWNOTE: Same basic story, except negative interaction now appears to be signif.


Subtitle	2SLS, Estimating CACE, Most Recent (Table A-7, top).
2SLS		EDImmigImp with ImmigRecentReceived, Female,NonWhiteH,
		Age,EducYrs,Income,PolInterest,Libcon,Rep/
		Instruments = ImmigRecent, Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon,Rep.
*ns.

Subtitle	2SLS, Estimating Interaction CACE (Table A-7, bottom).
Compute		Int = Rep*PredImmigRecent.
2SLS		EDImmigImp with ImmigRecentReceived, Female,NonWhiteH,
		Age,EducYrs,Income,PolInterest,Libcon,Rep,Int/
		Instruments = ImmigRecent, Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon,Rep,Int.
*NS, Int negaive.

*Comment        ****************************************.

Subtitle	Does Info Search Predict Perceived Importance? (Table A-2).
Regression	Variables = EDImmigImp,GHSImmigAtt,
		Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,Rep/Depend = EDImmigImp/Enter.
*Yes! Pos and Sig, p < .03.
Regression	Variables = EDImmigImp,GHImmigTOD,
		Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,Rep/Depend = EDImmigImp/Enter.
*Close, p < .19.

*Comment	****************************************
		Assumes Agenda-Setting = Stories read.

Subtitle	First Stage, Using Manip to Predict Discretionary Attention.
Regression 	Variables = GHSImmigAtt,ImmigCount, D1Immig,
	 	Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Rep/
		Depend = GHSImmigAtt/Enter/Save = Pred(PredImmigCountAtt).
Regression 	Variables = GHSImmigAtt,ImmigRecent, D1Immig,
	 	Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Rep/
		Depend = GHSImmigAtt/Enter/Save = Pred(PredImmigRecentAtt).
*NOTE: Instruments weak.
Descriptives	PredImmigCountAtt,PredImmigRecentAtt.

*Comment	****************************************.
*Comment	****************************************
		Assumes Agenda-Setting = Stories read.

Subtitle	Now 2SLS Again, Re-Estimating CACE.
2SLS		EDImmigImp with GHSImmigAtt,Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon,Rep/
		Instruments = ImmigCount, D1Immig,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Rep.
*neg, ns.

Compute		Int = Rep*PredImmigCountAtt.
2SLS		EDImmigImp with GHSImmigAtt,Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon,Rep,Int/
		Instruments = ImmigCount, D1Immig,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Rep,Int.
*NS, Int negative.

2SLS		EDImmigImp with GHSImmigAtt,Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon,Rep/
		Instruments = ImmigRecent, D1Immig,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Rep.
*Neg, ns.

Compute		Int = Rep*PredImmigRecentAtt.
2SLS		EDImmigImp with GHSImmigAtt,Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon,Rep,Int/
		Instruments = ImmigRecent, D1Immig,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Rep,Int.
*NS, int negative.


*Comment        **********************************************.
*Comment        **********************************************.

Title		FLAG BURNING AMENDMENT.
Subtitle	Flag Burning Amendment: Crucial Variables.
Descriptives	D1FlagBurn,EDFlagBurn,
		FlagBurnCount,FlagBurnDaysWt,FlagBurnRecent,
		FlagBurnDays,FlagBurnRecentReceived,
		FlagBurnOpens,FlagBurnUniqueOpens, 
		GHFlagBurnAtt,GHSFlagBurnAtt,GHFlagBurnTOD,
		D1FlagBurnImp,EDFlagBurnImp.

*Comment        **********************************************
		Does Assigned Treatment Affect Perceived Importance?.

Subtitle	Agenda-Setting, Flag Burning Amendment: ITT (Table A-3, top).
Regression	Variables = EDFlagBurnImp,FlagBurnCount,
		Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,Rep/Depend = EDFlagBurnImp/Enter.
*NS.

Subtitle	Flag Burning Amendment: ITT, Most Recent (Table A-3, bottom).
Regression	Variables = EDFlagBurnImp,FlagBurnRecent,
		Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,Rep/Depend = EDFlagBurnImp/Enter.
*NS, negative.

*Comment	*****************************************
		Assumes Agenda-Setting = Opportunity.

Subtitle        1st Stage, Using Manip to Predict Good ImmigDays (Table A-5).
Regression	Variables = FlagBurnDays,FlagBurnCount,
	 	Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Rep/
		Depend = FlagBurnDays/Enter/Save = Pred(PredFlagBurnCount).
Regression	Variables = FlagBurnRecentReceived,FlagBurnRecent,
	 	Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Rep/
		Depend = FlagBurnRecentReceived/Enter/
		Save = Pred(PredFlagBurnRecent).
*Note: Very good instruments!.
Variable Labels	PredFlagBurnCount "Instrument for FlagBurnDays"/
		PredFlagBurnRecent "Instrument for FlagBurnRecentReceived".
Descriptives	PredFlagBurnCount,PredFlagBurnRecent.

*Comment	*****************************************
		Assumes Agenda-Setting = Opportunity.

Subtitle	Now 2SLS, Estimating CACE (Table 1).
2SLS		EDFlagBurnImp with FlagBurnDays, Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon,Rep/Instruments = FlagBurnCount,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Rep.
*Neg, ns.

Subtitle	2SLS, Estimating Interaction CACE (Table 2).
Compute		Int = Rep*PredFlagBurnCount.
2SLS		EDFlagBurnImp with FlagBurnDays, Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon,Rep,Int/Instruments = FlagBurnCount,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Rep,Int.
*Neg ns, but Int pos & Sig, p < .05.

Subtitle	Separate Groups Procedure for Estimating Crucial Interaction.
temporary.
Select if	(Rep EQ 0).
2SLS		EDFlagBurnImp with FlagBurnDays, Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon/Instruments = FlagBurnCount,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon.
temporary.
Select if	(Rep EQ 1).
2SLS		EDFlagBurnImp with FlagBurnDays, Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon/Instruments = FlagBurnCount,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon.
*NEWNOTE: Same basic story, except lose signif because of smaller Ns.


Subtitle	2SLS, Estimating CACE, Most Recent (Table A-7, top).
2SLS		EDFlagBurnImp with FlagBurnRecentReceived, Female,NonWhiteH,
		Age,EducYrs,Income,PolInterest,Libcon,Rep/
		Instruments = FlagBurnRecent, Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon,Rep.
*neg, ns.

Subtitle	2SLS, Estimating Interaction CACE (Table A-7, Bottom).
Compute		Int = Rep*PredFlagBurnRecent.
2SLS		EDFlagBurnImp with FlagBurnRecentReceived, Female,NonWhiteH,
		Age,EducYrs,Income,PolInterest,Libcon,Rep,Int/
		Instruments = FlagBurnRecent, Female,NonWhiteH,Age,EducYrs,
		Income,PolInterest,Libcon,Rep,Int.
*Neg ns, but Int pos & Sig, p < .10.

*Comment        ****************************************.

Subtitle	Does Info Search Predict Perceived Importance? (Table A-2).
Regression	Variables = EDFlagBurnImp,GHSFlagBurnAtt,
		Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,Rep/Depend = EDFlagBurnImp/Enter.
*NS, but positive.

*Comment	****************************************
		Assumes Agenda-Setting = Stories read.

Subtitle	First Stage, Using Manip to Predict Discretionary Attention.
Regression 	Variables = GHSFlagBurnAtt,FlagBurnCount,
	 	Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Rep/
		Depend = GHSFlagBurnAtt/Enter/
		Save = Pred(PredFlagBurnCountAtt).
Regression 	Variables = GHSFlagBurnAtt,FlagBurnRecent,
	 	Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Rep/
		Depend = GHSFlagBurnAtt/Enter/
		Save = Pred(PredFlagBurnRecentAtt).
*NOTE: Instruments passable.
Descriptives	PredFlagBurnCountAtt,PredFlagBurnRecentAtt.

*Comment	****************************************
		Assumes Agenda-Setting = Stories read.

Subtitle	Now 2SLS Again, Re-Estimating CACE.
2SLS		EDFlagBurnImp with GHSFlagBurnAtt,Female,NonWhiteH,Age,
		EducYrs,Income,PolInterest,Libcon,Rep/
		Instruments = FlagBurnCount,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Rep.
*Neg, ns.

Compute		Int = Rep*PredFlagBurnCountAtt.
2SLS		EDFlagBurnImp with GHSFlagBurnAtt,Female,NonWhiteH,Age,
		EducYrs,Income,PolInterest,Libcon,Rep,Int/
		Instruments = FlagBurnCount,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Rep,Int.
*Neg ns, but Int pos & sig, p < .06.

2SLS		EDFlagBurnImp with GHSFlagBurnAtt,Female,NonWhiteH,Age,
		EducYrs,Income,PolInterest,Libcon,Rep/
		Instruments = FlagBurnRecent,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Rep.
*Neg ns.

Compute		Int = Rep*PredFlagBurnRecentAtt.
2SLS		EDFlagBurnImp with GHSFlagBurnAtt,Female,NonWhiteH,Age,
		EducYrs,Income,PolInterest,Libcon,Rep,Int/
		Instruments = FlagBurnRecent,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,Rep,Int.
*Neg ns, but Int pos & close, p < .12.






*Comment	************************************************.
 Title                        PRIMING.
*Comment	************************************************.
*For Reference: Candidates' General Ideology: 
		D Gov - lib, R Gov - con;
	        D Sen - strong lib, R Sen - mod;
		D House - mod, R House - strong con.

Subtitle	"Respondent's Policy Attitudes".
Descriptives    D1PFL,EDPFL,D1FlagBurn,EDFlagBurn,
		D1Immig,EDImmig,D1PolBrut,EDPolBrut.

Descriptives	PFLDGovDist,PFLDSenDist,PFLDHouseDist,
		PFLRGovDist,PFLRSenDist,PFLRHouseDist,
		PolBrutDGovDist,PolBrutDSenDist,PolBrutDHouseDist,
		PolBrutRGovDist,PolBrutRSenDist,PolBrutRHouseDist,
		FlagBurnDGovDist,FlagBurnDSenDist,FlagBurnDHouseDist,
		FlagBurnRGovDist,FlagBurnRSenDist,FlagBurnRHouseDist,
		ImmigDGovDist,ImmigDSenDist,ImmigDHouseDist,
		ImmigRGovDist,ImmigRSenDist,ImmigRHouseDist.
*NOTE: Disagreement is scored Hi, so all of these variables should be
       NEGATIVELY related to liking the different candidates.

Descriptives	Female,NonWhiteH,Age,EducYrs,Income,PID,Libcon,Dem,Rep,
                PolBrutCount,PredPolBrutCount,PolBrutCount3,D1PolBrut,
		PFLCount,PredPFLCount,PFLCount3,D1PFL,
                ImmigCount,PredImmigCount,ImmigCount3,D1Immig,
		FlagBurnCount,PredFlagBurnCount,FlagBurnCount3,D1FlagBurn.




*Comment	****************************************************.
*Comment	****************************************************.

Subtitle	Tests of the Priming Hypothesis, Dem Gov, ITT.
Compute		PFLInt = PFLCount*PFLDGovDist.
Compute		PolBrutInt = PolBrutCount*PolBrutDGovDist.
Regression	Variables = DemGovFT,PFLCount,PFLDGovDist,PFLInt,
			PolBrutDGovDist,PolBrutInt,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,PID/
		Depend = DemGovFT/Enter.
*OK, decent; Both interations negative, though neither signf.

Subtitle	Tests of the Priming Hypothesis, Dem House, ITT.
Compute		PFLInt = PFLCount*PFLDHouseDist.
Compute		PolBrutInt = PolBrutCount*PolBrutDHouseDist.
Regression	Variables = DemHouseFT,PFLCount,PFLDHouseDist,PFLInt,
			PolBrutDHouseDist,PolBrutInt,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,PID/
		Depend = DemHouseFT/Enter.
*Poor: 1 positive, 1 negative interaction, both small.


Subtitle	Tests of the Priming Hypothesis, Dem Gov, CACE Table 3.
*NOTE: Cannot do with SPSS's 2SLS program, because have two IVs.
Compute		PFLInt = PredPFLCount*PFLDGovDist.
Compute		PolBrutInt = PredPolBrutCount*PolBrutDGovDist.
Regression	Variables = DemGovFT,PredPFLCount,PFLDGovDist,PFLInt,
			PolBrutDGovDist,PolBrutInt,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,PID/
		Depend = DemGovFT/Enter.
*OK, decent; Both interations negative, though neither signf.

Subtitle	Tests of the Priming Hypothesis, Dem House, CACE Table 3.
Compute		PFLInt = PredPFLCount*PFLDHouseDist.
Compute		PolBrutInt = PredPolBrutCount*PolBrutDHouseDist.
Regression	Variables = DemHouseFT,PredPFLCount,PFLDHouseDist,PFLInt,
			PolBrutDHouseDist,PolBrutInt,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,PID/
		Depend = DemHouseFT/Enter.
*Poor: Both interactions positive, albeit small.

*Comment	****************************************************.

Subtitle	Tests of the Priming Hypothesis, Rep Gov, ITT.
Compute		FlagBurnInt = FlagBurnCount*FlagBurnRGovDist.
Compute		ImmigInt = ImmigCount*ImmigRGovDist.
Regression	Variables = RepGovFT,ImmigCount,ImmigRGovDist,ImmigInt,
			FlagBurnRGovDist,FlagBurnInt,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,PID/
		Depend = RepGovFT/Enter.
*Nada; 1 positive, 1 negative, neither signif.
Subtitle	Tests of the Priming Hypothesis, Rep House, ITT.
Compute		FlagBurnInt = FlagBurnCount*FlagBurnRHouseDist.
Compute		ImmigInt = ImmigCount*ImmigRHouseDist.
Regression	Variables = RepHouseFT,ImmigCount,ImmigRHouseDist,ImmigInt,
			FlagBurnRHouseDist,FlagBurnInt,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,PID/
		Depend = RepHouseFT/Enter.
*Bad: both positive, one almost significantly!.

Subtitle	Tests of the Priming Hypothesis, Rep Gov, CACE Table 3.
Compute		FlagBurnInt = PredFlagBurnCount*FlagBurnRGovDist.
Compute		ImmigInt = PredImmigCount*ImmigRGovDist.
Regression	Variables = RepGovFT,PredImmigCount,ImmigRGovDist,ImmigInt,
			FlagBurnRGovDist,FlagBurnInt,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,PID/
		Depend = RepGovFT/Enter.
*One negative iinteraction, one positive, neither significant.

Subtitle	Tests of the Priming Hypothesis, Rep House, CACE Table 3
Compute		FlagBurnInt = PredFlagBurnCount*FlagBurnRHouseDist.
Compute		ImmigInt = PredImmigCount*ImmigRHouseDist.
Regression	Variables = RepHouseFT,PredImmigCount,ImmigRHouseDist,
			ImmigInt,FlagBurnRHouseDist,FlagBurnInt,
		Female,NonWhiteH,Age,EducYrs,Income,PolInterest,Libcon,PID/
		Depend = RepHouseFT/Enter.
*Horrible: both positive, one signif.


*Comment	**********************************************.
*Comment	**********************************************.

Title    Does Saliency Moderate Priming Effect?.
Descriptives	Q_ED_LnDUR_Pol_FamLv,Q_ED_LnDUR_Pol_PlcBrut,
        	Q_ED_LnDUR_Pol_FlgBrn,Q_ED_LnDUR_Imp_Immig.

Subtitle        ITT.
Regression	Variables = Q_ED_LnDUR_Pol_PlcBrut,
		PolBrutRecent, Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,PID/
		Depend = Q_ED_LnDUR_Pol_PlcBrut/Enter.
Regression	Variables = Q_ED_LnDUR_Pol_FamLv
		PFLRecent, Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,PID/
		Depend = Q_ED_LnDUR_Pol_FamLv/Enter.

Regression	Variables = Q_ED_LnDUR_Imp_Immig,
		ImmigRecent, Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,PID/
		Depend = Q_ED_LnDUR_Imp_Immig/Enter.
Regression	Variables = Q_ED_LnDUR_Pol_FlgBrn
		FlagBurnRecent, Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,PID/
		Depend = Q_ED_LnDUR_Pol_FlgBrn/Enter.

Subtitle        CACE, Table A-6.
*NOTE: Since these are durations, recency predicts negative coefficients.
Regression	Variables = Q_ED_LnDUR_Pol_PlcBrut,
		PredPolBrutRecent, Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,PID/
		Depend = Q_ED_LnDUR_Pol_PlcBrut/Enter.
*Neg, but ns.
Regression	Variables = Q_ED_LnDUR_Pol_FamLv,
		PredPFLRecent, Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,PID/
		Depend = Q_ED_LnDUR_Pol_FamLv/Enter.
*Neg, but ns.

Regression	Variables = Q_ED_LnDUR_Imp_Immig,
		PredImmigRecent, Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,PID/
		Depend = Q_ED_LnDUR_Imp_Immig/Enter.
*Neg, but ns.
Regression	Variables = Q_ED_LnDUR_Pol_FlgBrn,
		PredFlagBurnRecent, Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,PID/
		Depend = Q_ED_LnDUR_Pol_FlgBrn/Enter.
*Neg, but ns.


*Comment	**********************************************.
*Comment	**********************************************.
*Comment	**********************************************.


Title		Reviewer #1, Point #2: Who Are Noncompliers?.
Subtitle	New Table A-4: Who are the Compliers (/Good Subjects)?.
Regression	Variables = ExpDaysGood,NonExpDaysGood,PolInterest,
		Newspaper,WebNews,TrustMedia,Libcon,PID,StrngPID,
		Age,Female,NonwhiteH,EducYrs,Income/Depend = ExpDaysGood/
		Enter =  Age,Female,NonwhiteH,EducYrs,Income/
		Enter = PolInterest,Newspaper,WebNews,TrustMedia,Libcon,StrngPID/
		Enter = NonExpDaysGood.


*Comment        ************************************************.

Title		"R#1, Point 6, Don't Assume Linearity of Treatment Effects".
*Frequencies	PredPolBrutCount,PredPFLCount,PredImmigCount,PredFlagBurnCount.
*Note: ,315 thru .695 is first dummy, .695 thru 1 is second dummy.
Recode		PredPolBrutCount (.3717 thru .4444 = 1)(Else = 0)
			into PredPolBrut2/
		PredPolBrutCount (.4444 thru hi = 1)(Else = 0)
			into PredPolBrut34/
		PredPFLCount (.3676 thru .4391 = 1)(Else = 0)
			into PredPFL2/
		PredPFLCount (.4391 thru hi = 1)(Else = 0)
			into PredPFL34/
		PredImmigCount (.2312 thru .4414 = 1)(Else = 0)
			into PredImmig2/
		PredImmigCount (.4414 thru hi = 1)(Else = 0)
			into PredImmig34/
		PredFlagBurnCount (.3632 thru .5652 = 1)(Else = 0)
			into PredFlagBurn2/
		PredFlagBurnCount (.5652 thru hi = 1)(Else = 0)
			into PredFlagBurn34.
Frequencies        PredPolBrut2,PredPolBrut34,PredPFL2,PredPFL34,
                PredImmig2,PredImmig34,PredFlagBurn2,PredFlagBurn34.
*Note: Good, distributions look appropriate.

Subtitle	Now 2SLS, Estimating CACE, Police Brutality.
Regression	Variables = EDPolBrutImp,PredPolBrut2,PredPolBrut34,
		Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,Dem/Depend = EDPolBrutImp/Enter.
Compute		Int2 = Dem*PredPolBrut2.
Compute		Int34 = Dem*PredPolBrut34.
Regression	Variables = EDPolBrutImp,PredPolBrut2,PredPolBrut34,
		Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,Dem,Int2,Int34/Depend = EDPolBrutImp/Enter.

Subtitle	Now 2SLS, Estimating CACE, Paid Family Leave.
Regression	Variables = EDPFLImp,PredPFL2,PredPFL34,
		Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,Dem/Depend = EDPFLImp/Enter.
Compute		Int2 = Dem*PredPFL2.
Compute		Int34 = Dem*PredPFL34.
Regression	Variables = EDPFLImp,PredPFL2,PredPFL34,
		Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,Dem,Int2,Int34/Depend = EDPFLImp/Enter.

Subtitle	Now 2SLS, Estimating CACE, Immigration Reform.
Regression	Variables = EDImmigImp,PredImmig2,PredImmig34,
		Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,Rep/Depend = EDImmigImp/Enter.
Compute		Int2 = Rep*PredImmig2.
Compute		Int34 = Rep*PredImmig34.
Regression	Variables = EDImmigImp,PredImmig2,PredImmig34,
		Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,Rep,Int2,Int34/Depend = EDImmigImp/Enter.

Subtitle	Now 2SLS, Estimating CACE, Flag Burning Amendment.
Regression	Variables = EDFlagBurnImp,PredFlagBurn2,PredFlagBurn34,
		Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,Rep/Depend = EDFlagBurnImp/Enter.
Compute		Int2 = Rep*PredFlagBurn2.
Compute		Int34 = Rep*PredFlagBurn34.
Regression	Variables = EDFlagBurnImp,PredFlagBurn2,PredFlagBurn34,
		Female,NonWhiteH,Age,EducYrs,Income,
		PolInterest,Libcon,Rep,Int2,Int34/Depend = EDFlagBurnImp/Enter.

*Comment        ************************************************.

Compute    Democrats = Dem.
Compute    Republicans = Rep.
Formats    Democrats,Republicans (F1.0).

Title		R#2, Point 9: Anova report df, F, eta, exact p.
Do Repeat	X = D1PFLImp,D1PolBrutImp,D1FlagBurnImp,D1ImmigImp.
Compute		X = 4*X + 3.
End Repeat.
Formats		D1PFLImp,D1PolBrutImp,D1FlagBurnImp,D1ImmigImp (F1.0).
Descriptives    D1PFLImp,D1PolBrutImp,D1FlagBurnImp,D1ImmigImp,
		EDPFLImp,EDPolBrutImp,EDFlagBurnImp,EDImmigImp.

Subtitle	Agenda-Setting, Paid Family Leave.
MANOVA		D1PFLImp,EDPFLImp by PFLCount3 (1,3) 
		Democrats (0,1)/WSFACTORS=Time(2)/
		Print = Signif(Efsize).

Subtitle	Agenda-Setting, Police Brutality.
MANOVA  	D1PolBrutImp,EDPolBrutImp by PolBrutCount3 (1,3) 
		Democrats (0,1)/WSFACTORS=Time(2)/
		Print = Signif(Efsize).

Subtitle	Agenda-Setting, Flag Burning Amendment.
MANOVA		D1FlagBurnImp,EDFlagBurnImp by FlagBurnCount3 (1,3) 
		Republicans (0,1)/WSFACTORS=Time(2)/
		Print = Signif(Efsize).

Subtitle	Agenda-Setting, Immigration Reform.
MANOVA		D1ImmigImp,EDImmigImp by ImmigCount3 (1,3) 
		Republicans (0,1)/WSFACTORS=Time(2)/
		Print = Signif(Efsize).

*Comment        ***********************************************
                NOW COMPLIERS ONLY.

Filter by Exp4DaysGood.
Subtitle	Agenda-Setting, Paid Family Leave, Compliers Only.
MANOVA		D1PFLImp,EDPFLImp by PFLCount3 (1,3) 
		Democrats (0,1)/WSFACTORS=Time(2)/
		Print = Signif(Efsize).

Subtitle	Agenda-Setting, Police Brutality, Compliers Only.
MANOVA  	D1PolBrutImp,EDPolBrutImp by PolBrutCount3 (1,3) 
		Democrats (0,1)/WSFACTORS=Time(2)/
		Print = Signif(Efsize).

Subtitle	Agenda-Setting, Flag Burning Amendment, Compliers Only.
MANOVA		D1FlagBurnImp,EDFlagBurnImp by FlagBurnCount3 (1,3) 
		Republicans (0,1)/WSFACTORS=Time(2)/
		Print = Signif(Efsize).

Subtitle	Agenda-Setting, Immigration Reform, Compliers Only.
MANOVA		D1ImmigImp,EDImmigImp by ImmigCount3 (1,3) 
		Republicans (0,1)/WSFACTORS=Time(2)/
		Print = Signif(Efsize).
Filter off.
